基于冲激响应加窗法的CO-OFDM系统信道估计研究

Channel estimation for CO-OFDM system based on impulse response processed with window

针对相干光正交频分复用(CO-OFDM)系统中光纤的色散和信号传输过程中噪声对系统可靠性的降低,提出了信道的冲激响应加窗(IRM,impulse response processed with window)算法,在最小二乘(LS)算法的基础上,通过时域加窗将信道冲激响应长度以外的噪声滤除,保证了信道冲激响应长度在OFDM保护间隔之内。在算法复杂度提高并不大的前提下,IRW算法系统误码率(BER)比LS算法降低了近1个数量级。仿真结果表明,在256个子载波的CO-OFDM系统中,BER为10-4时,IRW算法对系统信噪比(SNR)的要求比LS算法低了近2.5dB,而只比线性最小均方误差(LMMSE)算法高0.7dB;并且,IRW算法的复乘次数仅比LS算法高8倍,而比LMMSE算法的复乘次数低32倍。

To reduce the influence of fiber dispersion and noise in coherent optical orthogonal frequency division multiplexing（CO-OFDM） system,we present a new channel estimation method based on impulse response processed with window（IRW） algorithm.This method relies on least square（LS） estimation,and then uses the channel impulse response processed with the time domain window to filter the noise outside the window to ensure the length of channel impulse response in OFDM guard interval.The IRW method is not more complex than LS algorithm,and has lower bit error rate（BER） than LS algorithm by nearly an order of magnitude.The simulation results show that in the 256-subcarrier CO-OFDM system,and with the same BER of 10-4 the required signal-to-noise ratio（SNR） in IRW algorithm is lower than that in the LS algorithm by nearly 2.5 dB,while only 0.7 dB higher than that in the linear minimum mean square error（LMMSE） algorithm.At the same time,the number of complex multiplying of IRW algorithm is only 8 times higher than that of LS,but nearly 32 times lower than that of LMMSE.